You might be misunderstanding the difference between short-term forecasts and longterm projections. I know I failed to understand the scientific nuance until recently.
You see, "global average temperatures are going to rise X by 2100" is a projection. It's based on pretty basic thermodynamics (ie. this much carbon increases the greenhouse effect by such-and-such). This science, because it's so basic, is pretty solid.
At the same time "global average temperatures are going to rise by Y by 2025" is a forecast. It's based on computer models that are perpetually being refined to more accurately predict the short-term trend. Most recently, these models were found to be missing el-nino/la-nina cycles which is why they have lagged over the last decade.
This is why people get confused when I tell them the science of global warming is actually extremely basic. It's just thermodynamics, but then they confuse projections with forecasts and wonder why the models haven't accurately predicted the last 10 years. It's the "weather versus climate" debate all over again.
Why do scientists even publish forecasts when they know they are still very much a work in progress? Politics. You see, your local representative couldn't give a damn if your children's children suffer from today's lack of leadership a century from now. So scientists are tasked to find out what the short-term effect will be on the constituency to inform politicians whether or not they might suffer some voter backlash on the issue.
In other words, our children's children are doomed to shell out billions to fix this mess.